The Stimulatory Adenosine Receptor ADORA2B Regulates Serotonin (5-HT) Synthesis and Release in Oxygen-Depleted EC Cells in Inflammatory Bowel Disease

Objective We recently demonstrated that hypoxia, a key feature of IBD, increases enterochromaffin (EC) cell 5-HT secretion, which is also physiologically regulated by the ADORA2B mechanoreceptor. Since hypoxia is associated with increased extracellular adenosine, we wanted to examine whether this nucleotide amplifies HIF-1α-mediated 5-HT secretion. Design The effects of hypoxia were studied on IBD mucosa, isolated IBD-EC cells, isolated normal EC cells and the EC cell tumor derived cell line KRJ-1. Hypoxia (0.5% O2) was compared to NECA (adenosine agonist), MRS1754 (ADORA2B receptor antagonist) and SCH442146 (ADORA2A antagonist) on HIF signaling and 5-HT secretion. Antisense approaches were used to mechanistically evaluate EC cells in vitro. PCR and western blot were used to analyze transcript and protein levels of HIF-1α signaling and neuroendocrine cell function. An animal model of colitis was evaluated to confirm hypoxia:adenosine signaling in vivo. Results HIF-1α is upregulated in IBD mucosa and IBD-EC cells, the majority (∼90%) of which express an activated phenotype in situ. Hypoxia stimulated 5-HT release maximally at 30 mins, an effect amplified by NECA and selectively inhibited by MRS1754, through phosphorylation of TPH-1 and activation of VMAT-1. Transient transfection with Renilla luciferase under hypoxia transcriptional response element (HRE) control identified that ADORA2B activated HIF-1α signaling under hypoxic conditions. Additional signaling pathways associated with hypoxia:adenosine included MAP kinase and CREB. Antisense approaches mechanistically confirmed that ADORA2B signaling was linked to these pathways and 5-HT release under hypoxic conditions. Hypoxia:adenosine activation which could be reversed by 5′-ASA treatment was confirmed in a TNBS-model. Conclusion Hypoxia induced 5-HT synthesis and secretion is amplified by ADORA2B signaling via MAPK/CREB and TPH-1 activation. Targeting ADORA2s may decrease EC cell 5-HT production and secretion in IBD.


Introduction
Inflammatory Bowel Disease (IBD) is highly prevalent in Europe and North America and a recent systematic review demonstrated an increasing incidence (for UC: 6.3-24.3/100,000; for CD: 5-20.2) [1]. This coupled with the long duration of the illness make IBD one of the most common gastroenterological diseases with a prevalence per 100,000 of 505 and 249 for UC and 322 and 319 for CD in Europe and the US, respectively [1]. The etiology and pathogenesis of IBD, however, remains largely unknown. While defects in local immune responses (both innate as well as adaptive) to commensal microflora and food antigens are assumed to play pathogenic roles in IBD [2,3], recent studies have also demonstrated a role for the enterochromaffin (EC) cell in the pathogenesis of this disease.
The EC cell is the most common neuroendocrine cell in the epithelia lining the lumen of the gut and plays a key regulatory role in gut secretion, motility, pain, and nausea [4]. The monoamine neurotransmitter serotonin (5-hydroxytryptamine: 5-HT) has proven central in EC cell regulatory function and these cells synthesize, store, and release the vast majority (95%) of the body's store of this amine [5]. EC cells function as ''taste buds of the gut'' and represent sensory transducers responding to mechanical events, luminal acidification, or nutrients such as glucose and short chain fatty acids, bile salt, tastants and olfactants [6][7][8][9][10][11][12][13]. In addition, EC cell secretion can be activated by neural, bacterial and immunological input [14,15]. Specifically, development of IBD is associated with altered EC cell serotonin release [15,16].
Serotonin is considered to play a role in IBD through activation of immune cell types which express receptors for this amine [15,17]. TPH-1 knockout mice respond to chemically-induced colitic agents with a less severe phenotype and delayed onset of disease compared to wild-type mice treated in the same protocol [15]. A variety of other studies [18][19][20] support a role for serotonin in modulating immune signaling and the promotion of interactions between innate and adaptive immune responses within the context of gut inflammation.
Recently, rhythmic mechanical strain that mimics normal bowel movements (mediated by ADORA2B receptors) has been identified to induce EC cell secretion and transcription of EC cell secretory products -responses that are accentuated by neoplasia [21]. We have also demonstrated that gut EC cells are oxygenresponsive and alterations in O 2 levels differentially activate HIF-1a signaling and serotonin release [22]. This results in alterations in serotonin production and secretion, effects amplified by inflammation. In addition, to the latter, alterations in neuroendocrine signaling as well as activation of hypoxia-mediated responses are features recently identified in a TNBS animal model [23] and in IBD samples through transcriptome analyses [24].
Hypoxia is also strongly associated with an increase in extracellular/mucosal adenosine levels [25] and with stabilization of HIF-1a [26]. HIF-1a induces transcription and increases the activity of 59ecto-nucleotidase (CD73), the enzyme that converts AMP to adenosine [27]. CD73 also regulates transcription of the ADORA2B receptor while suppressing transcription of the adenosine re-uptake transporters, equilibrative nucleoside transporters 1 and 2 (ENT1 and 2). Furthermore, CD73 decreases the intracellular metabolism of adenosine by suppressing the transcription of adenosine kinase [28]. In IBD, localized hypoxia occurs as a result of chronic inflammation increasing the metabolic needs of the tissue [29], and thus potentially up-regulates the adenosine-ADORA pathway. ADORA2B is the predominant ADORA receptor in colonic mucosa [30] and is also up-regulated by TNFa [31]. Activation of the receptor is thought to regulate cytokine production including IL-10 [32]; colitis is reduced in knockout mice [33,34] suggesting a protective role.
We hypothesized that the increase in 5-HT observed in IBD may, in part, be due to hypoxia increasing functional HIF-1a which triggers an increase in extracellular adenosine signaling, leading to increased production and secretion of 5-HT via ADORA2B receptor activation. Gut mucosal tissue from IBD patients, isolated EC cells and the well-characterized EC cell line KRJ-1 were studied. This cell line possesses similar properties (e.g. similar signaling pathways, enzyme activity and secretory products) and have similar responsiveness to stimuli, as normal EC cells and is therefore an appropriate model to study 5-HT regulation [13,21,35,36].

Human Samples
Tissue was collected from twenty-one patients (M:F = 12:9; median age [range] = 53 yr ). CD tissue (n = 12) was obtained from patients who had undergone surgery for CD ileitis (n = 3) or colitis (n = 9). Only grossly affected tissue was studied. Macroscopically ''normal'' tissue was obtained from matched samples when available (n = 6). All tissue was collected between 2008 and 2013 at Yale University, Department of Surgery following written informed consent from patients per protocol (Yale University School of Medicine IRB approval, HIC#0805003870).

ADORA2B Knockdown
A 19-mer oligonucleotide antisense corresponding to 461-480 of the rat ADORA2B receptor (NM_01716.1) was designed to induce a steric obstacle for protein translation (Yale Medical School Keck Oligonucleotide Synthesis Facility) [38]. Control nucleotides were prepared with randomized sequence of matching nucleotides per protocol. In these experiments, isolated EC cells [39] were exposed to oligonucleotides (antisensense: TCCCTCTTGCTCGTGTTCC, or control: CTGTTCCGTCCGTTCCCTT -150 pmol) for 12 hrs (FITCuptake of oligonucleotides was noted as early as 2 hrs within cells, with peri-nuclear uptake complete by 14-16 hrs), and then assessed for mRNA, flow cytometry (receptor expression), secretion and by western blot. These experiments were conducted within 16 hrs following oligonucleotide uptake.

EC Cell Isolation
EC cells (.98% purity) were isolated from human or rat samples by mucosal stripping, enzymatic digestion, and a combination of Nycodenz gradient fractionation and fluorescence activated cell sorting (FACS) as described [13,16,35,39]. Approximately 1610 6 cells were obtained per mucosal sample, a quantity sufficient for real-time PCR, short-term culture and western blots.

Cell Culture Studies
KRJ-I cells [40] were maintained as floating aggregates in Quantum 263 complete tumor growth medium (PAA) supplemented with penicillin (100 IU/ml) and streptomycin (100 ug/ ml). EC cells (normal, IBD or isolated from rat) were maintained in short-term culture (,12 hrs after isolation) under the same conditions. All experiments were performed without antibiotics; the cell line was mycoplasma free.
Hypoxic conditions were induced using a modular incubator chamber (MIC-101, Billups-Rothenberg Inc, Del Mar, CA). Briefly, short-term cultured EC cells or cultured KRJ-I cells (48 hrs) were transferred to the humidified hypoxic chamber; the chamber was flushed with CO 2 for 4 min to maintain hypoxic conditions (0.5% O 2 ). KRJ-I cells (4610 5 cells/ml, n = 6) were seeded in 6 well plates (Falcon, BD, Franklin Lakes, NJ) NECA, curcumin, SCH442146, MRS1754, and DMSO were added to the wells. DMSO was added to the controls to compensate for NECA and MRS1754 being solubilized in DMSO (,0.1% final concentration). Cells were then incubated for 15 minutes. They were then exposed to hypoxia for 0, 15, 30, 60, 120 and 240 mins.
After cells were harvested, whole-cell lysates were prepared by adding 200 ml of ice-cold cell lysis buffer ( ). Tubes were centrifuged at 12,000 g for 20 min and protein amount in the supernatant was quantified using the BCA protein assay kit (Thermo Fisher Scientific, Rockford, IL).

Serotonin Secretion
5-HT levels were analyzed using commercially available ELISA assays (5-HT: BA 10-0900; Rocky Mountain Diagnostics) as previously described [13] in supernatant according to the manufacturer's instructions.

RLU Studies
The Cignal HIF Pathway Reporter Assay Kit (LUC) (CCS-007L) was used to evaluate HIF signaling in EC cells (human, ADORA2B-antisense treated rat) and in KRJ-I cells. Briefly, the basis of this protocol is transient transfection with a HIFresponsive luciferase construct that encodes the firefly luciferase reporter gene under the control of a minimal (m)CMV promoter and tandem repeats of the hypoxia transcriptional response element (HRE). This is designed to monitor the activity of HIFregulated signal transduction pathways in cultured cells. Each reporter is premixed with constitutively expressing Renilla luciferase, which serves as an internal control for normalizing transfection efficiencies and monitoring cell viability. Short-term cultured CD EC cells (10,000/well) or KRJ-I cells (10,000/well) were transfected per protocol. Renilla luciferase activation following ADORA2 activation was measured using the dual luciferase assay (Glomax). The average maximum response per kit is 4 RLU; in these experiments 2 RLU were identified.

Western Blot Analysis
Analyses were performed on 30 min hypoxia samples to evaluate total TPH, p-TPH, total CREB, p-CREB, total ERK, p-ERK, HIF-1a and PKA. Total protein lysates (20 mg) were denaturated in SDS sample buffer, separated on a Tris-Glycine gel (10%) and transferred to an Immobilon P (PVDF) membrane (Milipore Corporation, Bedford, MA). After blocking (5% BSA for 60 min at room temperature) the membrane was incubated with primary antibodies (Cell Signaling Technology and BD Biosciences) in 5% BSA/PBS/Tween20 overnight at 4uC. The membranes were incubated with the horseradish peroxidaseconjugated secondary antibodies (Cell Signaling Technology) for 60 min at room temperature and immunodetection was performed using the Western Lightning TM Plus-ECL (PerkinElmer, MA). Blots were exposed on X-OMAT-AR films [40,41]. The optical density of the appropriately sized bands was measured using ImageJ software (NIH, USA).

RNA Isolation, Reverse Transcription and RT-PCR Analyses
RNA was extracted from mucosa (macroscopically normal human or rat -normal, TNBS and TNBS-treated with 59ASA), isolated normal and CD EC cells (1610 6 ), isolated normal and ADORA2B-deficient (antisense) rat EC cells, and KRJ-I cells (1610 6 , n = 4-7) using TRIZOLH (Invitrogen, Carlsbad, CA) then cleaned (Qiagen, RNeasy kit, Qiagen, Valencia, CA) and converted to cDNA (High Capacity cDNA Archive Kit, Applied Biosystems, Carlsbad, CA) [36,42]. RT-PCR analyses were performed using Assays-on Demand TM and the ABI 7900 Sequence Detection System [36,42]. Primer sets (HIF-1a (human and rat) and ADORA2B (rat)) were all obtained from Applied Biosystems and PCR mix on gels were performed to confirm presence of single bands for each primer set. PCR Data was normalized using the DDC T method; ALG9 was used as a housekeeping gene [43] for human, GAPDH was used for rat [23].

Flow Cytometry
Rat EC cells (control or antisense at 12 hrs) were stained with ADORA2B (1:500) and flow cytometry was conducted on a BD FACS Aria Cell Sorter (BD Biosciences, Bedford MA). Positive cells were identified for unstained, and the two EC cell populations (antisense-treated and control).

Statistics
Results were expressed as mean6standard error (SEM). All statistical analyses were performed using Prism 4 (GraphPad Software, San Diego, CA). Results were compared between control and stimulated cells using the Mann-Whitney test. A p,0.05 was considered significant.

HIF-1a Expression in Normal and IBD Mucosa and EC Cells
HIF-1a transcripts were increased 3.5-fold in IBD mucosa compared to macroscopically normal CD mucosa ( Figure 1A). A similar pattern was evident at the EC cell level, with CD EC cells demonstrating a ,2.5-fold increase of HIF-1a mRNA compared to normal EC cells. Assessment of protein expression confirmed HIF-1a activation both in mucosa as well as in EC cells isolated from Crohn's mucosa ( Figure 1B). Immunofluorescent staining of CgA and HIF-1a ( Figure 1C) identified double positive cells in the mucosa (yellow arrows), confirming that the HIF-1a-positive cells were enteroendocrine. Significantly more EC cells (,90%, p,0.001) were positive in Crohn's mucosa than in macroscopically normal mucosa ( Figure 1D). These results suggest that EC cells are exposed to hypoxia during inflammation and predominantly exhibit an activated hypoxia-mediated signaling pathway (HIF-1a).

The 5-HT Secretory Pathway
To evaluate whether adenosine-mediated HIF1a activation increased 5-HT release from EC cells, we examined the effects of hypoxia with and without ADORA2 antagonists on KRJ-I cells over a 4 hr time period. Hypoxia significantly increased 5-HT between 15 and 120 mins with a maximal effect (2.2 fold, p,0.05 versus no hypoxia, Figure 2A). Curcumin, a known HIF-1a inhibitor through transcriptional repression [46], reversed hypoxia-mediated secretion at all time points. NECA, increased 5-HT release (30-120 min), while MRS1754 but not SCH442146, decreased it (15-120 min). This suggests that 5-HT release by hypoxia is driven, at least in part, by activation of ADORA2B receptors, and that adenosine can modulate 5-HT secretion.
We next evaluated expression of enzymes involved in 5-HT synthesis and vesicle uptake (TPH-1 and VMAT-1) and in secretion per se, chromogranin A (CgA). We focused on 30 mins as this identified the time point at which 5-HT was maximally secreted. Total protein levels of TPH-1 were unchanged by hypoxia at 30 min ( Figure 2B) but the phosphorylated form of this enzyme, which identifies activated TPH-1, was increased. This was amplified by NECA and inhibited by MRS1754. Analysis of the ratio of activated to total TPH-1 protein identified that this was increased 1.660.16 by hypoxia (p,0.05 vs. controls) and 2.460.2 (by NECA: p,0.05 vs. hypoxia) and was reduced to 0.9160.17 (by MRS, p,0.05 vs. hypoxia). VMAT-1 functions to accumulate cytosolic monoamines, like 5-HT, into secretory vesicles [47]. No significant changes were noted for VMAT-1 by hypoxia, but levels were significantly elevated by NECA (2.8360.07, p,0.05) ( Figure 2C), indicating that adenosine signaling activates 5-HT uptake into vesicles. CgA is important for granulogenesis and  secretion in neuroendocrine cells [48]. No alterations were identified ( Figure 2D). We interpret this to indicate that hypoxia and adenosine directly activate pathways associated with the formation of components essential to neuroendocrine secretion.
To evaluate HIF-1a mediated signaling we undertook transient transfection with a HIF-responsive firefly luciferase construct under HRE-transcription control (Renilla luciferase-constructs) in KRJ-I cells ( Figure 3B) and in IBD-EC cells ( Figure 3C). In KRJ-I cells NECA activated luciferase (RLU) under normoxic conditions while MRS1754 inhibited this; no effect was noted for SCH442146. Under hypoxic conditions, NECA amplified luciferase activity (3-fold, p,0.05 vs. hypoxia) which was inhibited by MRS1754 but not by SCH442146. In IBD-EC cells (which have an activated HIF-1a- Figure 1A-C), activation of HRE-mediated transcription was amplified by NECA (,1.5-fold RLU) under normoxic conditions and inhibited by MRS1754. These results suggest that adenosine, similar to a reduction in O 2 , can increase HIF-1a protein levels and induce HRE-signaling which is mediated via the ADORA type 2B receptor.
CREB is an important transcription factor for TPH-1 transcription [50][51][52]. While not significantly increased by hypoxia, the phosphorylated form was significantly increased by NECA ( Figure 4B), an effect that was reversed by MRS1754. This suggests a role for ADORA-mediated signaling in the regulation of hypoxia-mediated TPH1 transcription.
PKA, which mediates the adenosine signal for 5-HT production and secretion under mechanical stress [21], was not significantly changed by hypoxia and was not altered by either NECA or MRS1754 ( Figure 4C) suggesting that this pathway does not regulate 5-HT secretion under these conditions.

Mechanistic Analysis of Hypoxia-adenosine Signaling Pathways in EC Cells
We used an antisense approach to mechanistically dissect the pathways associated with hypoxia-adenosine signaling in isolated rat EC cells. ADORA2B antisense decreased transcription (,50%) and was associated with a significant and almost complete reduction in ADORA2B membrane protein expression ( Figure 5A). These cells could respond to hypoxia with 5-HT release, but this was significantly lower than cells with normal ADORA2B expression ( Figure 5B). The latter cells responded similar to human EC cells with adenosine-regulated hypoxiamediated 5-HT release, while the antisense-treated cells had largely lost adenosine mediated effects. At a protein level, we identified reduced pMAPK, pCREB and pTPH-1 in antisense treated hypoxic cells ( Figure 5C) identifying that these are the principle pathways involved in adenosine-mediated 5-HT secretion.

Hypoxia and ADORA2B in an Animal TNBS-induced Colitis Model
Finally, we evaluated expression of HIF-1a and ADORA2B in intestinal mucosa from a rat animal TNBS-induced colitis model and evaluated the effect of 5-ASA. A real-time PCR analysis confirmed significant upregulation of HIF-1a and ADORA2B mRNA by TNBS ( Figure 6A). The expression levels were decreased by 5-ASA treatment. Protein levels followed a similar expression pattern ( Figure 6B). We interpret these data to confirm activation of a hypoxia-adenosine pathway in a colitis model -similar to observations in clinical samples.

Discussion
IBD is associated with abnormalities in the pro-inflammatory 5-HT system. Previously, we have demonstrated that LPS and IL1b induce a significantly elevated 5-HT response in IBD mucosa [16]. This may partly explain the hypersecretion of 5-HT noted in the IBD disease process. As IBD mucosa is also exposed to significant hypoxic stress, we examined the effects of hypoxia on EC cell 5-HT secretion. We have identified that hypoxia induces 5-HT secretion from EC cells and that this can be reversed by targeting HIF-1a. HIF-1a is also associated with increased mucosal adenosine availability, a known regulator of 5-HT production and secretion [22]. Based on these observations, we investigated the relationship between hypoxia, the ADORA signaling system and 5-HT production/secretion in EC cells from CD and normal mucosa to evaluate whether these factors are involved in the amplified 5-HT production noted in IBD.
We demonstrated that HIF-1a was increased both at the mRNA and protein levels in IBD mucosa, in an animal model of colitis and in IBD-EC cells. The latter almost predominantly (,90%) expressed a hypoxia activated phenotype and active HRE signaling. These data confirm that the inflamed mucosa is under hypoxic stress [53,54], and that EC cells, in particular, are hypoxia-activated. We have previously identified increased ADORA2B receptors both at protein and transcript levels in EC cells isolated from IBD patients compared to controls [21]. In the current study, we identified that adenosine amplified HIF-1a expression and activity, which suggests that it can act as a positive feedback mechanism for 5-HT synthesis and release.
Activating ADORA signaling via NECA increased 5-HT release which could specifically be inhibited by MRS1754 identifying that, under low oxygen mucosal levels such as found with hypoxia, EC cell secretion is regulated by adenosine and specifically, via activation of ADORA2B receptors. The inhibitory effect of MRS1754 was complete for the time period 15-90 mins but at 120 mins, we could identify no inhibition by this agent. This suggests that the adenosine (ADORA2B):hypoxic signaling pathway occurs early (within 30 mins) and that other ADORA receptors may play a role later in hypoxia-mediated 5-HT release; it is unlikely that this is ADORA2A. A combination of HIF-1a signal activation and adenosine-mediated ADORA2 signaling appears to play important roles in EC cell 5-HT secretion in IBD.
HIF-1a itself may directly regulate transcription of the ratelimiting enzyme in 5-HT synthesis TPH-1. The promoter region of this gene encodes hypoxia responsive elements (HRE) [55]; hypoxia may therefore directly increase 5-HT production by increasing TPH-1 expression. Transcriptional alterations in TPH-1 were not consistently identified in our current study (data not shown) suggesting the major route regulating 5-HT release occurred at the level of protein regulation and via pathways cross-activated by HIF-1a signaling.   We have previously demonstrated that the adenosine/ ADORA2B/cAMP/PKA/CREB pathway plays a pivotal role in regulating 5-HT secretion from the EC cell when subject to mechanical stress [21]. In the current study, we demonstrated that the activation of the ADORA2B receptor was important for the increased levels of 5-HT noted during hypoxia. The intracellular signaling pathways involved in the hypoxic response, however, exhibit differences to that identified for EC cell mechanoresponsivity. We failed to detect differences in PKA suggesting that activation of cAMP-responsive pathways is not PKAregulated under these conditions. In contrast, we identified ADORA2B mediated activation of MAPK signaling as well as increased phosphorylation (and thereby activation) of TPH-1. TPH-1 is a MAPK target and it is likely, under hypoxic conditions, that MAPK phosphorylates and thereby activates this enzyme, leading to an increase in 5-HT synthesis. Antisense approaches mechanistically confirmed roles for both MAPK as well as TPH-1 in EC cells under hypoxic conditions.
Other factors involved in 5-HT secretion were altered by adenosine-ADORA2B signaling. Specifically, protein levels of the rate-limiting enzyme involved in 5-HT vesicular accumulation, VMAT1, were increased. We postulate that the excess 5-HT produced by activated TPH-1, is actively transported into vesicles which then provides a large pool for release. We interpret this to demonstrate that vesicle maturation [47,48], is directly regulated by adenosine under hypoxic conditions.
Adenosine plays a complex role in IBD. Apart from activating EC cell 5-HT secretion, this nucleotide can also decrease SERT activity [56]; the combination resulting in a increased mucosal 5-HT signal. One prediction, given the pro-inflammatory activity of 5-HT, is an exacerbation of colitis under these conditions. Other studies have demonstrated an increased severity of colitis in ADORA2B2/2 mice suggesting that targeting this receptor may be a potential therapeutic target. Currently, loss of ADORA2B is considered to result in down-regulation of IL10 production, with accentuation of colitis. Our studies indicate that loss of ADORA2B decreases 5-HT and, presumably, the mucosal aminergic signal.
Elevated mucosal adenosine, induced in hypoxic conditions or during abnormalities in peristalsis, may activate EC cell 5-HT synthesis and release, and reduce enterocyte-mediated SERT production with a resultant overall increase in 5-HT proinflammatory signaling. In the context of ADORA signaling, our studies and others, highlight both the complexity of the mucosal pathways involved in IBD as well as the importance of delineating individual cell signaling. This may help better direct targeted therapy.